Saphira is a web application that generates scientific hypotheses based on biological literature and visualizes their support vs. complexity using a percolation plot. It uses a backend powered by Express and SQLite, with a Next.js frontend for user interaction.
- Clone the repository:
git clone <repository-url> cd saphira cd backend
npm install npm run build npm run start
cd frontend npm install npm run dev
Ensure data/abstracts.json contains valid scientific abstracts.
The backend automatically loads and processes this data on startup.
Access the Application: Open http://localhost:3000 to view the frontend.
The backend runs on http://localhost:3001.
Hypothesis Generation: Generates hypotheses at complexity levels 2-5 based on literature.
-
Percolation Plot: Visualizes support rate vs. complexity using Plotly.
-
Complexity Slider: Allows users to select desired hypothesis complexity.
-
Literature Processing: Extracts entities and relations from abstracts using NLP.
Integrate Solana blockchain for decentralized hypothesis storage.
Enhance NLP with advanced models for better entity and relation extraction.
Add user authentication and hypothesis saving functionality.